That’s the discussion in this new IZA working paper from Halla and Schneider:
While there is an extensive literature on tax evasion a further aspect of cheating on the state, namely benefit fraud, has gained relatively modest attention in the economic literature. This paper seeks to fill this gap. We explore differences between benefit fraud and tax evasion due to differing social norms. We define the concepts of benefit morale and tax morale as the motivation to abstain from cheating on the state via these two offenses. Our multilevel analysis, based on a large micro data set of respondents from 29 OECD member countries, shows that benefit morale and tax morale have different determinants at an individual-level and respond differently to fiscal policy measures.
So what do they find? Moral norms matter:
Our results suggest that moral values evolve endogenously and are determined by prices (i.e. the cost of acting morally). Citizens who have comparably more opportunities and low cost to commit a certain offense, develop the attitude that it is a minor offense. This suggests that citizens excuse or rationalize their own deviant behavior. Put differently, they self-servingly adjust their moral values. From a policy perspective it is beneficial to know which groups of citizens view benefit fraud as a minor offense and which sub-population is reluctant to pay taxes. This allows policy makers to predict how cheating behavior will evolve over time in response to socio-demographic changes.
But incentives matter too, because they shape moral norms:
We have shown that in the context of benefit fraud and tax evasion the intrinsic motivation to comply (benefit morale and tax morale) is indeed altered by extrinsic factors, such as tax rates.